Mexico - Housing Market Development Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No.: 19095 IMPLEMENTATION COMPLETION REPORT THE UNITED MEXICAN STATES HOUSING MARKET DEVELOPMENT PROJECT (LOAN 3497-ME) March29, 1999 Mexico Country Management Unit Finance, Private Sector and Infrastructure Sectors Unit Latin America and Caribbean Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank Authorization.  CURRENCY EQUIVALENTS (as of March 29, 1999) Currency Unit = Mexican Peso ($) {Currency} = US${Amount} US$1.00 = MexNP 3.09 average of 1992 MexNP 3.11 average of 1993 = MexNP 3.39 average of 1994 = MexNP 6.50 average of 1995 = MexNP 7.59 average of 1996 = MexNP 7.95 average of 1997 = MexNP 9.06 average of 1998 Monthly minimum wage (in the Federal District) 1992 = MexNP 399.9 (May, 1992) 1993 = MexNP 428.1 (average) 1994 = MexNP 458.1 (average) 1995 = MexNP 538.8 (average) 1996 = MexNP 669.0 (average) 1997 = MexNP 793.5 (average) WEIGHTS AND MEASURES One Meter = 3.38 feet (ft) FISCAL YEAR January 1 - December 31 ABBREVIATIONS AND ACRONYMS BANOBRAS - Banco Nacional de Obras y Servicios Puiblicos S.N.C. (National Development Bank for Public Works and Services) CETES - Certificados de la Tesoreria de la Federaci6n (Treasury Bills) CPP - Costo Promedio Porcentual de Captacion (Average Cost of Funds to the Banking System) FICAPRO - Fideicomiso Casa Propia (Fund for Home Ownership) FONHAPO - Fondo de Habitaciones Populares (Govermment's Low-Income Housing Fund) FOVI - Fondo de Operaci6n y Financiamiento Bancario a la Vivienda (Government Housing Finance Fund) FOVISSSTE - Fondo de la Vivienda del Sistema de Seguridad Social de los Trabajadores del Estado (Housing Fund for Public Sector Workers) GDP - Gross Domestic Product INFONAVIT - Instituto del Fondo Nacional de la Vivienda para los Trabajadores (Institute of the National Housing Fund for Private Sector Workers) MMWs - Multiples of Monthly Minimum Wage SEDESOL - Secretaria de Desarrollo Social (Ministry of Social Development ) SHCP - Secretaria de Hacienda y Credito Publico (Ministry of Finance and Public Credit) SOFOLES - Sociedades Financieras de Objeto Limitado (Limited Purpose Financial Institutions) Vice President: Shahid Javed Burki Director Country Management Unit: Olivier Lafourcade Director Sector Management Unit: Daniel Leipziger Task Manager: Richard Clifford/ Sabino Escobedo  FOR OFICIAL USE ONLY TABLE OF CONTENTS Page PREFACE ................. . ........................... . ............. iii EVALUATION SUMMARY v................................................. iV PART I: PROJECT IMPLEMENTATION ASSESSMENT .......1....... A. Introduction ..................... .............................. 1 B. Project Objectives ...............................................1..... C. Achievement of Objectives ................................................ 2 D. Major Factors Affecting the Project ................ ..................... 5 E. Project Sustainability ................................................... 6 F. Bank Performance .................................................... 6 G. Borrower Performance ..................................................... 7 H. Assessment of Outcome ....................... .................. 8 I. Future Operation ......................................... 8 J. Key Lessons Leamned ................. ........................ 9 PART Il: STATISTICAL TABLES ........................ ....................... 10 Table 1: Summary of Assessment ......................................... 10 Table 2: Related Bank Loans/Credits ....................................... l 1 Table 3: Project Timetable ..................... .................... 12 Table 4: Loan/Credit Disbursements: Cumulative Estimated and Actual 13 Table 5A: Key Indicators for Project Implementation ........ ............ 14 Table 51B: Key Indicators for Project Implementation ........ ............ 15 Table 5C: Key Indicators for Project Implementation ........ ............ 16 Table 6A: Key Indicators for Project Operation .............. ............. 17 Table 6B: Key Indicators for Project Operation .............. ............. 18 Table 7: Studies Included in Project ........................................ 19 Table 8A: Project Costs ................. ........................ 20 Table 8B: Project Financing .................................................. 20 Table 9: Economic Costs and Benefits ............... ..................... 21 Table 10: Status of Legal Covenants ........................................ 22 Table 11: Bank Resources: Staff Inputs . .................................. 24 Table 12: Bank Resources: Missions ................ ...................... 25 A]PPENDIX-ES: A. Mission's Aide Memoire .................... 26 B. Borrower Contribution to the ICR .................... 28 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  iii IMPLEMENTATION COMPLETION REPORT MEXICO HOUSING MARKET DEVELOPMENT PROJECT (LOAN 3497-ME) PREFACE This is the Implementation Completion Report (ICR) for the Housing Market Development Project in Mexico for which Loan 3497-ME in the amount of US$ 450.0 million equivalent was approved on June 25, 1992 and made effective on November 23, 1992. The Loan was closed on December 31, 1997. It was fully disbursed and the last disbursement took place on May 20, 1997. The ICR was prepared by Mr. Howard A. Jones, consultant and reviewed by Richard Clifford, Task Manager and Country Sector Leader, Mexico Country Management Unit and by Maria Victoria Lister, Quality Assurance Officer (LCSFP). The borrower and cofinanciers provided comments that are included as appendixes to the ICR. Preparation of this ICR was begun during the Bank"s final supervision/completion mission, October 23, 1998. It is based on material in the project file. The borrower contributed to preparation of the ICR by contributing to the views expressed in the ICR and by preparing its own evaluation of project execution, which is attached.   v IMPLEMENTATION COMPLETICON REPORT MEXICO (Loan 3497-ME) EVALUATION SUMMARY INTRODUCTION 1. The project was the fifth in a series of projects supported by the Bank in the housing sector. The cost of the Housing Market Development Project was estimated to be US$1,594.3 million with a Bank loan of US$450 million equivalent. Final project costs were US$2,527.8 million, with US$450 million disbursed from the Bank loan. PROJECT OBJECTIVES 2. According to the Staff Appraisal Report (SAR) and the Loan Agreement (LA), project objectives were to: (i) expand the supply of funding for low-cost housing in Mexico; (ii) induce commercial banks to increase lending to creditworthy lower-income home buyers; (iii) support the removal of unnecessary regulatory barriers to promote the participation of financiers and the private developers in the low-cost housing sector of Mexico; (iv) improve the publicly-available information regarding the housing finance and housing production market in Mexico; (v) improve the wage index on which the Fondo de Operaci6n y Financiamiento Bancario a la Vivienda (FOVI) subloan repayment adjustments are based; (vi) to achieve the aims of the Program which, in addition to the project objectives set forth above would expand and liberalize the urban land market; improve the cost recovery and operational perfornance of the quasi-public pension funds' housing programs, increase the integration of the housing finance market with the overall financial system and promote investment in rental housing. 3. Through the achievement of these objectives, the project would increase the supply of affordable housing to lower income households. Improvements in the efficiency of the housing finance system and deregulation measures should be reflected in  vi a reduction in the relative price of housing and an increase in the housing production. The project studies would provide the analytical basis for important future government action to improve both public and private sector participation in the housing market and for the integration of the housing finance system with the overall financial market which should lead to enhanced liquidity and resource mobilization for mortgage lenders. The financing for low-and moderate-income households would help avoid the environmental problems derived from substandard housing. IMPLEMENTATION EXPERIENCE AND RESULTS 4. FOVI's low-income housing program includes four different types of housing units: (a) Type A, with maximum property value of 100 monthly minimum wages (mmw); (b) Type Bi, with maximum property value of 130 mmw; (c) Type B2, with maximum property value of 160 mmw; and (d) Type B3, with maximum property value of 190 mmw. Middle-income housing is considered to be housing units costing more than 190 mmw. 5. Because of a shift in demand to higher cost housing as housing costs rose relative to family income, the number of housing units costing up to 130 mmw, and financed by the Bank, totaled only 73,204 compared with 99,142 estimated at appraisal. The number of housing units costing over 130 mmw but less than 160 mmw financed by FOVI totaled 90,487 compared with 29,934 estimated at appraisal. Accordingly, the total project cost was significantly higher than the SAR estimate (US$2.5 billion compared with US$1.6 billion). In any event, neither FOVI nor the Bank financed middle-income housing units using the Loan funds. 6. Efforts to remove regulatory barriers and regulatory costs to promote the participation of financiers and developers and to reduce housing costs in the low-cost housing sector were highly successful and fell from an average of 11.7% of housing costs in 1992 to 4.7% in 1997. Publicly-available infonnation on housing finance and production was improved and made available on a monthly basis to subscribers. An improved wage index to adjust housing loan repayments was developed and tested but has not been put into use for fear of its market impact. A study of the secondary mortgage market was carried out and produced useful conclusions which have been taken into account in designing the new FOVI Restructuring Project (LN 4443-ME). 7. Although the Constitution of Mexico was amended in 1992 to permit the transfer of ejido land, the increase in land made available for housing has been much less than expected due to difficulties in establishing ownership rights to occupied land. However, the Secretaria de Desarrollo Social (SEDESOL) has been working with state and local governments to accelerate these land transfers. While organizational reformns have been made to both the Instituto del Fondo Nacional de la Vivienda para los Trabajadores (INFONAVIT) and the Fondo de la Vivienda del Sistema de Seguridad Social del los  vii Trabajadores del Estado (FOVISSSTE) and their operations are more market based and more efficient, both institutions continue to provide large housing subsidies. The Fondo de Habitaciones Populares (FONHAPO) was also fundarmentally restructured and now operates more effectively through state housing institutes rather than social groups which were unable to account for and manage housing loans. Rent controls have now been largely eliminated but efforts to promote rental housing have had limited results as credit auctions favor blocks of new owner-occupied housing rather than isolated units. 8. Despite the effects of the 1994 and subsequent (levaluations, FOVI assets grew somewhat more than SAR estimates as a significantly higher number of houses were financed during project execution. Income was higher than appraisal estimates due to higher interest rates and financial spreads on a larger number of operations. Also FOVI's debt to equity ratio was slightly better than the SAR estimate. Although FOVI now guarantees 50% of subloans made by financial intermedLiaries, its financial condition is sound. 9. The two major factors affecting project execution were: (i) a significant increase in housing costs relative to increases in the minimum wage; and (ii) the impact of the economic crisis in 1994-1995 which seriously and adversely affected affordability and the availability of resources for housing finance. FOV][ reacted imaginatively to these problems by offering a partial guarantee on subloans, lowering repayment levels and extending repayment terms from 20 years to 25 and then 30 years and increasing bank commissions. 10. Bank performance in project identification, preparation, appraisal and supervision was satisfactory throughout. Performance of the borrower was highly satisfactory during project preparation and appraisal. FOVI and the government provided a great deal of useful information in preparing the project and worked closely with Bank staff in the preparation of the project and in a detailed sector policy statement with was incorporated into the project. Borrower performance during project imaplementation was satisfactory. SUMMARY OF FINDINGS, FUTURE OPERATION, AND KEY LESSONS LEARNED 11. Project implementation was satisfactory for most of the project's duration and project objectives were either fully or partially achieved. No internal rate of return on the investments was calculated. It was concluded that housing investments financed under the project were provided at the least cost taking into account FOVI's well-designed and competitive auction system for housing credits, its afforcdability criteria and the generally high level of competition amongst developers for financing. 12. Future operations depend upon plans to continue to expand and improve the operational efficiency and influence of FOVI in channeling finance for low-cost housing through the private sector in Mexico. For such housing, financial institutions will  viii continue to be heavily dependent on FOVI financing until the development of a secondary mortgage market is realized. The FOVI Restructuring Project will provide for (i) an increase of the flow of funds to the housing sector; (ii) a restructuring program that will strengthen FOVI's products and services for the primary and secondary markets, its finances and management; and (iii) the implementation of a new subsidy policy. 13. Despite the factors affecting the project described above, FOVI has continued to lend through the banks and has rapidly expanded its lending through the Sociedades Financieras de Objeto Limitado (SOFOLES). As described above, FOVI reacted imaginatively to adverse developments to sustain affordability and lending levels and has maintained a relatively strong financial condition with government support. FOVI is now in the process of expanding its staff and modernizing its systems in preparation for the next stage of its development with Bank support. 14. The key lessons learned relate to the importance of a close relationship between project design, ownership and the macro-economic climate. While there was a relatively large number of project objectives, they reflected a coherent whole contained in the policy statement of the government. Project ownership by the implementing agency (FOVI) and the government was essential as many of the objectives of the project depended not only on government support but on direct government action. The government was closely involved with the implementing agency in project preparation and in the development of a policy statement for the sector which strengthened ownership. It not only approved the plans for achieving regulatory and sector reforms but was actively involved in promoting reforms at the state and municipal levels. The macroeconomic climate was also important for project success. Rising housing costs relative to household income and the significant reduction in the availability of credit and its increased cost due to the 1995 economic crisis had serious adverse effects on the credit component. These risks were not foreseen during appraisal. Creative action by FOVI and the government ameliorated their effects. 15. Overall the project was successful. The problems encountered during implementation and the studies undertaken under the project led to a clearer understanding of the need for greater private sources of long-term finance for social housing.  1 IMPLEMENTATION COMPLETION REPORT MEXICO HOUSING MARKET DEVELOPMENT PROJECT (Loan 3497-ME) PART I: IMPLEMENTATION ASSESSMENT A. Introduction 1. The project was the fifth in a series of projects supported by the Bank in the housing sector. The cost of the Housing Market Development Project was estimated to be US$1,594.3 million with a Bank loan of US$450 million equivalent. Final project costs were US$2,527.8 million, with US$450 million disbursed from the Bank loan. B. Project Objectives 2. The objectives of the project were: (i) to expand the supply of mortgage finance for low-cost housing; (ii) to induce commercial banks to increase lending to creditworthy lower-income home buyers; (iii) to support the removal of unnecessary regulatory barriers to promote the participation of financiers and private developers in the low cost housing sector; (iv) to improve publicly-available information regarding the housing finance and housing production market in Mexico; (v) to improve the wage index on which FOVI subloan repayment adjustments are based; (vi) to achieve the aims of the Program which, in aLddition to the project objectives set forth above would expand and liberalize the urban land market; improve the cost recovery and operation performance of the quasi-public pension funds' housing programs, increase the integration of the housing finance market with the overall financial system and promote investment in rental housing. 3. In order to achieve these objectives, the project had four components: (i) a regulatory reform component; (ii) a housing sector policy component; (iii) a FOVI credit line component; and (iv) a studies component. Bank financing was provided only for the credit line component; the cost of the studies, which was expected to be small, was borne by FOVI and the Bank of Mexico. The SAR reflected an expectation that achievement of project objectives would assist the government in creating a more supportive environmnent in which the private sector would play an expanded role in solving the country's housing problems by: establishing a supportive legal and regulatory  2 environment; rationalizing public sector intervention in housing activities to facilitate the operation of market-based housing finance institutions; ensuring that key inputs for housing production are supplied, particularly land, infrastructure and financing; and providing adequate information so that policy makers and market participants can make sound decisions. 4. The regulatory reform component was designed to achieve objective (iii) of the project; the housing sector policy component to achieve objectives (ii) and (vi); the FOVI credit line to achieve objective (i) and the studies component to achieve objectives (iv), (v) and (vi). The policy component comprised the implementation of policies to: induce commercial banks to actively participate in financing housing for the middle-and low- income population with their own resources and those of FOVI; liberate the urban land market; modernize the housing funds supported by the compulsory savings system, integrate their operations with the rest of the financial markets and improve their cost recovery; improve the cost recovery of mortgages financed by FONHAPO; deregulate housing development, construction and titling at the state and municipal level; promote higher investment in rental housing; and improve the reliability and congruity of information on the housing market under a self-financing scheme. 5. The objectives of the project were clear and consistent with the project components. They were important for the country and its housing and financial sectors and for the Bank's country assistance strategy. Project objectives were generally realistic. However, policy goals to liberate the urban land market and integrate the housing funds with the operations of the rest of the financial system were ambitious and only partly achieved during project implementation. Project objectives were numerous and required the active support of the Bank of Mexico and the federal and state and local governments. The Bank of Mexico and the government provided the support to implement its policy statement and the project and to engage support at the state and local levels. 6. In order to minimize risks, the project included limits on loan disbursements to ensure adequate time for deregulation to be implemented and produce results. However, the main project risks were not that deregulation would not occur but that a crisis in financial markets would interfere with the achievement of housing finance goals. C. Achievement of Objectives 7. FOVI's low-income housing program includes four different types of housing units: (a) Type A, with maximum property value of 100 mmw; (b) Type B 1, with maximum property value of 130 mmw; (c) Type B2, with maximum property value of 160 mmw; and (d) Type B3, with maximum property value of 190 mmw. Middle- income housing is considered to be housing units costing more than 190 mmw. 8. The number of housing units financed by FOVI during the project totaled 163,691 compared to the 129,076 estimated in the SAR. However, units costing up to 130  3 monthly minimumn wages (mmw) and financed by the Bank, totaled only 73,204 compared with 99,142 estimated at appraisal. Because of the shift in demand to higher cost housing as housing costs rose significantly relative to family income, the number of housing units costing more than 130 mmw but less than 160 mmw totaled 90,487 during project implementation compared with 29,934 estimated at appraisal (the closing date for the loan was extended by a year until December 31, 1997 to complete the project although the final loan disbursement was made May 20, 1997). Accordingly, the total project cost was significantly higher than the SAR estinmate (US$2.5 billion compared with US$1.6 billion) and the Bank financed only 17.8% of project cost compared with 28% estimated at appraisal. Nevertheless, despite the economic crisis of 1995, FOVI's total housing financed significantly exceeded its lending targets for the project period. 9. Because of the shift in demand for finance to higher cost housing, the distribution of final credits costing up to 130 mmw fell short of the 70% target except for the year 1992 (Table 5A). The distribution of auction financing, however, achieved the target of 80% of credits for housing costing up to 130mnmw except in the years 1993 and 1994 when they were well short of that target (39% and 42% respectively). In order to maintain lending for lower cost housing in particular and maintain bank lending for housing in general, FOVI implemented a number of changes in its lending policies in 1993 and 1994. It established a partial guarantee system for credits (initially 55% or 60% of outstanding loan amounts, subsequently reduced to 50%), allowed banks to stop repaying FOVI for subloans which were more than four months overdue, reduced repayment factors on subloans and increased maximum repayment periods from 20 years to 25 years and then to 30 years to increase affordability, and increased commission's paid to banks and reduced bank capital backing for housing subloans by 50% to improve profitability. FOVI also changed the interest rates on subloans from the average cost of banks funds to a rate which was 5% in real terms. 10. Commercial bank lending for housing with their own and FOVI funds increased from some US$3.2 billion in 1991 to over US$5 billion annually in the years 1992 through 1994. However, lending with their own funds collapsed in 1995 when the economic crisis caused their non-performing loans to increase greatly and real interest rates to rise to very high levels. Housing credits financed with FOVI resources auctioned earlier were basically the only bank lending for housing during 1995 and 1996. To an increasing extent bank lending for housing has been taken over by SOFOLES (specialized financing institutions) which began to participate in FOVI housing programs in 1995. In 1997 SOFOLES accounted for 47% of FOVI lending up from 1% in 1995 and 18% in 1996. In the early 1990's independent companies, which had financed automobiles since the 1970's, began to form groups to finance houses (autofinanciamientos). Members of the groups make monthly payments to a fund and lotteries are held periodically to finance home purchases. They grew significantly after 1995 with the drop-off in commercial bank financing. 11. Efforts to remove unnecessary regulatory barriers to promote the participation of financiers and developers in the low-cost housing sector were highly successful. A project objective was for 22 states to sign deregulation agreements with FOVI during project implementation. By agreement, in October 1992 governors of all 31 states and the DDF (Federal District) signed agreements to deregulate and reduce taxes in the  4 housing sector and agreements were signed with the National Association of Notaries to reduce their fees to 1% of transaction value. As a result of these agreements, regulatory costs fell from an average of 11.7% of housing costs in 1992 to 4.7% in 1997 (Table 5B). States with lower regulatory costs also fared better in the FOVI auction process, as lower cost housing gave promoters a significant advantage in the bidding process. In 1993 cities with high regulatory costs (over 12%) received only 0.4% of credit rights; cities with costs between 8% and 12% received 33.6% and cities with costs less than 8% received 66%. 12. Efforts to improve the publicly-available information on housing finance and production were successful. SEDESOL now collects data from official and private sources and makes them available to subscribers on a monthly basis. It also issues annual reports with complete housing sector data. 13. A study to develop an improved wage index to adjust housing loan repayments was carried out by the Bank of Mexico and the Social Security Administration. The index developed for the manufacturing sector was subjected to a number of tests over the years and was found useful for measuring affordability; however, the index has not been used for fear of the negative effect its application could have on the demand for housing credit. 14. In 1992, the constitution was amended to allow transfers of ejido land (communal holdings) to individuals. It was expected that this would result in a rapid increase in the supply of urban land. However, the issuance of land ownership certificates has been very slow as it has been difficult to establish legal ownership rights to occupied land. A large amount of land has been made available for housing during project implementation . SEDESOL continues to work with state and federal authorities to accelerate the transfer of ejido land for housing developments. 15. During project implementation steps were taken to improve the operational performance of INFONAVIT and FOVISSTE, the quasi-public pension fund housing programs. INFONAVIT reformed its procedures for authorizing housing credits including taking into account the savings record of potential beneficiaries as a qualification for making a housing loan. FOVISSTE is also reorganizing its operations. Despite these reforms, however, both institutions continue to provide large housing subsidies, FOVISSTE through fixed nominal interest rates and INFONAVIT through subsidized rates, poor credit standards and weak loan servicing. 16. FONHAPO was also fundamentally restructured during project implementation. It closed all of its regional delegations, reduced its staff by some 60%, and now works exclusively through state housing institutions which are responsible for the promotion, control and recovery of housing credits made with FONHAPO resources. Previously, FONHAPO had made loans to social groups which were n1ot capable of accounting for and managing housing loans. 17. Also, and in order to increase competition for housing finance, the Secretary of Finance and Public Credit authorized the new financial institutions, the SOFOLES, to have access to FOVI credits for housing finance.  5 18. Efforts to promote investmnent in rental housing have had limited results. Although rent controls, which were a serious obstacle to investment in rental housing, have been eliminated in most areas, developers continue to perceive risks in the legal system, especially regarding tenant rights . 19. In order to prepare for further integration of the housing finance system into the rest of the financial market when long-term sources of savings become substantial, a study of the potential development of a secondary mortgage market was carried out under the project by the Bank of Mexico. The study was completed after a series of visits to other countries with secondary mortgage markets. The study produced useful conclusions for the development of such a market in Mexico, including the need to make any subsidy transparent and fully front-loaded and the interest rates on housing loans market based. 20. Financial Performance. FOVI assets grew faster than SAR forecasts (in US dollar equivalents) until 1995, as the number and value of houses financed was greater than estimated. With the large devaluation at the end of 1994 and further devaluation in 1995 assets declined but grew subsequently and by 1997 exceeded SAR estimates (Table 6A). Income was much higher than estimated at appraisal due to higher interest rates and spreads and a larger number of houses financed than estimated. Administrative expenses were close to SAR estimates. Also, FOVI's debt equity ratio was slightly better than the SAR estimate (16.1% compared to 13.9% ) at the end of 1996. 21. Capitalized interest (negative amortization) is now close to 50% of total mortgage loans outstanding compared with 24% at appraisal. This is due to the very high interest rates in 1995 and 1996. However, maximum loan terms are now 30 years rather than 20 years at appraisal. FOVI has carried out a number of simulations of loan repayment scenarios and concluded that any unpaid balances at the end of the 30 year term would be very small (if any) unless the macroeconomic imbalances which occurred in 1995 were repeated. Although FOVI guarantees up to 50% of subloans, no commercial bank has exercised its guarantee rights. In 1995 rather than foreclosing loans delinquent housing loans were largely refinanced with government support. Due to the high demand for housing even in weak economic periods, losses on the exercise of any guarantees are expected to be relatively small and the losses manageable by FOVI. D. Major Factors Affecting the Project 22. The two major factors affecting the execution of the project were: (i) a significant increase in housing costs relative to increases in the minirmum wage; and (ii) the impact of the economic crisis of 1995 which seriously and adversely affected affordability and the availability of financial resources for housing finance. In the first case FOVI took a number of measures as set out in para. 9 to sustain lending and maintain affordability. In the second case, rising unemployment, declining real wages, and very high real interest rates led to a serious deterioration in the commercial banks' loan portfolios in general and their housing loan portfolios in particular. This resulted in a virtual cessation of new  6 bank lending for housing and FOVI suspended its housing for eight months at the beginning of 1995. The government assisted the banks in restructuring their loans to help strengthen the financial system generally. E. Project Sustainability 23. Despite the factors affecting the project described above, FOVI has continued to lend through the banks and has rapidly expanded its lending through the SOFOLES. As described above FOVI reacted imaginatively to adverse developments to sustain affordability and lending levels and has maintained a relatively strong financial condition with government support. FOVI is now in the process of expanding its staff and modernizing its systems in preparation for the next stage of its development with Bank support. The objectives of a new Bank loan for which FOVI will be executing agency will be to provide a renewal of the flow of funds to the sector, implementation of a new housing subsidy policy and a strengthening of FOVI's products and services for the primary and secondary markets. F. Bank Performance 24. Project Identification and Preparation: Satisfactory. The project was identified in early 1991, appraised in early 1992, approved by the Board in June 1992 and became effective in November 1992 (Table 3). From identification to appraisal, project issues were progressively defined through a series of three preparation missions in April, September and November 1991. The project had a high priority with the government and grew out of extensive Bank experience with Mexico's housing sector. It was consistent with the Bank's strategy for Mexico to increase private sector participation in housing finance while continuing to liberalize and integrate financial markets. 25. A Bank team with expertise in the range of areas dealt with in the project (urban, financial, secondary mortgage markets, environmental) worked closely with FOVI staff in project preparation and received important inputs from FOVI and other government staff. During project preparation a revision of the government's housing sector policy was carried out based on the National Housing Program for 1990-1994. The project was developed within this policy context and government plans to reform the financial sector, including privatization, deregulation, and strengthening of the quasi-public housing funds. 26. Project Appraisal: Satisfactory. The commitment of the government and FOVI were confirmed during appraisal and the government's housing policy statement reviewed. The loan amount was reviewed and considered adequate for the items to be financed with Bank support during the project implementation period. The implementation plan and performance indicators were sound for FOVI operations, but the quality of data on commercial bank and public sector lending for housing and information on the urban land market proved to be poor. Macroeconomic risks to the  7 project were not recognized. The Bank relied upon the imanagement capacity of FOVI with which it had had considerable experience, and FOVI, with strong support from Bank of Mexico specialized staff, responded well to problems as they emerged during project implementation. 27. Supervision: Satisfactory. The continuity of essential Bank staff involved in project preparation and appraisal was maintained for the first two years of project implementation. The Task Manager for the project was changed in 1994 when the mid- term review was carried out and subsequently project supervision was assumed increasingly by the Bank's Resident Mission staff. Two supervision missions were scheduled in each of the years 1992 and 1993. Close cooperation was maintained between the Bank and FOVI which provided detailed annual reports on project implementation along lines which had been agreed with the Bank. The Bank provided important assistance to the government in arranging visits to a number of foreign countries to discuss their housing programs for the poorest and their secondary mortgage markets. The participation of specialized Bank staff was arranged to assist FOVI to evaluate the impact of its new loan guarantee program. Subsequent supervision missions in 1995 and 1996 identified with FOVI the need for a new housing loan and a grant to finance its preparation. Based upon a request by FOVI, Bank staff also supported the extension of the closing date for the loan by a year in view of the hiatus in FOVI lending in 1995 following the crisis. G. Borrower Performance 28. Preparation and Appraisal: Highly Satisfactory. The Bank had worked with FOVI since the mid-1980s and the preparation, appraisal and implementation of the Housing Finance Project which was successfully carried out and fully disbursed in 1992 nearly two full years ahead of the appraisal estimate. In the preparation and appraisal of the Housing Market Development Project, FOVI worked closely with the Bank team and supplied it detailed data on low-cost housing credit demand, financing by FOVI and the quasi-public housing agencies, financial sector issues, land issues, regulatory issues, other sector issues and policies, the costs and terms of new housing products to be delivered, proposed studies, etc. FOVI also carried out a number of simulations of loan recovery rates under various economic scenarios. Work with the Bank team was close and fruitful and project preparation and appraisal were efficient and expeditious. Government commitment to the project goals was well coordinated amongst FOVI, the Bank of Mexico, SEDESOL and the private sector. 29. Implementation: Satisfactory. Despite the economic crisis in 1995 and its serious adverse effects on the supply of credit for low-income housing, FOVI managed to maintain fairly high levels of lending throughout project implementation. Initially, as the demand for housing shifted to higher cost units as housing price increases continued to outpace increases in family income, FOVI took a number of steps to address the problem as spelled out in para. 9 above. It also took a series of measures in 1996 to make housing more affordable to beneficiaries and more profitable to financial institutions. FOVI has been quite imaginative in its efforts to maintain lending to the sector. Efforts of the federal government to promote the reduction in regulatory costs at the state and local  8 level were highly successful as were efforts to improve information available to the private sector on housing production and finance. Project studies were completed and have been useful in defining the aforementioned information requirements but also in understanding the modifications in lending design needed to develop a secondary mortgage market. While significant improvements were made to FONHIAPO operations and institutional arrangements, and some improvement were made in the lending of INFONAVIT and FOVISSTE subsidies provided through these two institutions continue to be quite high. Also, despite the significant step taken to allow the sale of ejido land, progress in increasing available land for housing has been slow. Nevertheless, overall, considering the range of issues addressed in the project, project implementation was judged to be satisfactory. 30. Compliance with Loan Covenants: Satisfactory. The status of implementation of the major legal covenants is given in Table 10. Compliance was generally good except for a continuing problem in obtaining complete audited information on project costs and sources of finance. H. Assessment of Outcome 31. Overall Rating: Satisfactory. Implementation was satisfactory for most of the project's duration and the project objectives were either fully or partially achieved. No internal rate of return on the investments was calculated. It was concluded that housing investments financed under the project were provided at the least cost taking into account FOVI's well-designed and competitive auction system for housing credits, its affordability criteria and a generally high level of competition amongst developers for financing. Efforts to reduce regulatory barriers and costs also provided incentives to lend in least cost states and municipalities. I. Future Operation 32. Future operations depend upon plans to continue to expand and improve the operational efficiency and influence of FOVI in channeling finance for low-cost (social interest) housing through the private sector in Mexico. When the project was completed, FOVI was financially sound and lending through a number of financial agents. Lending through the commercial banks is recovering and lending through the SOFOLES expanding. However, for social interest housing, financial institutions will continue to be heavily dependent on FOVI financing until the development of a secondary mortgage market is realized. Until this occurs the supply of credit for social housing will probably continue to fall well short of need and the deficit for such housing will continue to grow. Accordingly, the FOVI Restructuring Project will provide for (i) a renewal of the flow of funds to the housing sector; (ii) a restructuring program that will strengthen FOVI's products and services for the primary and secondary markets, its finances and management; and (iii) the implementation of a new subsidy policy that will remove subsidies from the mortgage instrument.  9 J. Key Lessons Learned 33. The lessons learned and discussed below underscore the importance of close linkages between project design, ownership and the macro--economic climate. 34. Project Design. While there was a relatively large number of project objectives, they reflected a coherent whole contained in the sector policy statement of the government. The project had four components, and while the credit component was an important one, implementation of the regulatory reform and housing sector policy components was essential to the long-term success of FOVI's credit operations. The studies were also important to the success of FOVI's operations and the future shift of low-cost housing finance from public institutions to private long-term sources of funding. Project design also took into account the Bank's positive experience with FOVI, and the strong support FOVI had received from the Bank of Mexico and the overall experience the Bank had in housing operations in Mexico. 35. Project Ownership. Project ownership by the implementing agency FOVI and the government was essential as many of the objectives of the project depended not only on government support but on direct government involvement. The government was involved with FOVI in project preparation and in the development of a policy statement for the sector which strengthened ownership. The govermnent not only approved the plans for achieving regulatory and sector reforms but was actively involved in promoting reforms at the state and municipal levels. 36. Macroeconomic Climate. This was also very important for project success. The shortfalls in the credit component were largely due to the inflation of housing costs which adversely affected affordability, and the crisis of 1995 which significantly reduced the availability of credit for housing and greatly increased its cost. These risks were not foreseen at appraisal. However, the actions of FOVI and the government to a significant degree ameliorated their effects on FOVI operations. 37. Despite the important achievement of revising the Constitution to remove restrictions on the sale of ejido land, the amount of land made available for housing development from this source during project execution was less than expected at appraisal. The Bank underestimated the obstacles in the way of privatizing this land and making it available for housing. 38. Overall the project was successful although there were some shortfalls in the implementation of the credit line, in the availability of land for housing and in the institutional strengthening of the quasi-public housing agencies. However the problems encountered in the implementation of the credit component and the problems which emerged in relying so heavily on the banking system, led to a clearer understanding of the need for other private sources of finance for social housing. This understanding was advanced by the study included in the project on the secondary mortgage market and on further sector work carried out in preparation of a follow-up project with FOVI. FOVI's experience and the studies pointed to a clear need for redesigning the housing credit instrument in order to be able to tap pension funds and contractual savings through secondary market operations.  10 PART II: STATISTICAL TABLES Table 1: Summary of Assessments A. Achievement of Objectives Substantial Partial Negligible Not applicable (0) (0) (0) (v) Macro Policies [1 * a O Sector Policies 0 O O